by , , ,
Abstract:
How can one find a subset, ideally as small as possible, that well represents a massive dataset? I.e., its corresponding utility, measured according to a suitable utility function, should be comparable to that of the whole dataset. In this paper, we formalize this challenge as a submodular cover problem. Here, the utility is assumed to exhibit submodularity, a natural diminishing returns condition preva- lent in many data summarization applications. The classical greedy algorithm is known to provide solutions with logarithmic approximation guarantees compared to the optimum solution. However, this sequential, centralized approach is imprac- tical for truly large-scale problems. In this work, we develop the first distributed algorithm – DISCOVER – for submodular set cover that is easily implementable using MapReduce-style computations. We theoretically analyze our approach, and present approximation guarantees for the solutions returned by DISCOVER. We also study a natural trade-off between the communication cost and the num- ber of rounds required to obtain such a solution. In our extensive experiments, we demonstrate the effectiveness of our approach on several applications, includ- ing active set selection, exemplar based clustering, and vertex cover on tens of millions of data points using Spark.
Reference:
Distributed Submodular Cover: Succinctly Summarizing Massive Data B. Mirzasoleiman, A. Karbasi, A. Badanidiyuru, A. KrauseIn Proc. Neural Information Processing Systems (NeurIPS), 2015Spotlight presentation
Bibtex Entry:
@inproceedings{mirzasoleiman15cover,
	author = {Baharan Mirzasoleiman and Amin Karbasi and Ashwinkumar Badanidiyuru and Andreas Krause},
	booktitle = {Proc. Neural Information Processing Systems (NeurIPS)},
	month = {December},
	title = {Distributed Submodular Cover: Succinctly Summarizing Massive Data},
	year = {2015}}